Introduction
Postoperative pulmonary complications (PPCs) adversely influence the mortality and hospital stays of patients after upper abdominal surgery [
1,
2]. Their incidence after abdominal surgery ranges from 10 to 80% [
3,
4]. Hepatectomy is one kind of common upper abdominal surgery, performed close to the diaphragm, and patients undergo this surgery often complicated with liver insufficiency and hypoproteinemia [
5], these lead to higher incidence of PPCs in this kind of patients. Meanwhile, with the development of critical care medicine and the increase quantity of surgeries, more and more critical patients are transferred to the ICU after hepatectomy. These patients have more complex complications and more likely to develop the PPCs [
6,
7].
Several studies reporting the risk factors of pulmonary complications after hepatectomy, including a few multivariate analyses, are usually limited in those non-critical patients [
4,
8,
9]. A small number of studies have looked at critically ill populations, but these studies have small sample sizes and just included few variables, and most do not contain detailed information about intraoperative anesthesia or surgery [
5,
10].
To the best of our knowledge, there have been no high-quality studies on the risk of pulmonary complications after hepatectomy in the critically ill population and no study have created a nomogram for prediction of the PPCs in these people. And the analysis of the factors related to the occurrence of PPCs in critical patients and build a prediction model will have great clinical value in establishing a feasible path for clinical intervention to reduce the occurrence of PPCs in this kind of patients. In this study, data from a 7-year clinical database were retrospectively analyzed, aimed to determine the risk factors for PPCs of critical patients after hepatectomy and build a nomogram model for the prediction of PPCs.
Discussion
Several studies have reported on PPCs after hepatic surgery. However, to the best of our knowledge, no high-quality multivariate analysis has been published on a group of critical patients after hepatectomy in a large population and create a nomogram model for prediction of PPCs. And in our study, we use a clear and rigorous definition to define the PPCs [
13].
In this retrospective study, we observed that PPCs occurred in 55.7% of the critical patients admitted in the ICU after hepatectomy. The rates of overall pleural effusion, respiratory failure, atelectasis, respiratory infection, pneumothorax, aspiration pneumonitis, and bronchospasm were 42.5% (n = 214), 31.7% (n = 159), 22.5% (n = 113), 11.1% (n = 56), 0.4% (n = 2), 0.2% (n = 1), and 0% (n = 0), respectively. The incidence of PPCs in our study was higher than that in previous studies [
8,
16,
17], which is due to the fact that all patients in our study were critical, and their overall situation after surgery was more complicated. The presence of PPCs was significantly associated with increased postoperative mortality, longer mechanical ventilation time, and prolonged length of stay (LOS) in the ICU and hospital.
In this study, we identified four independent risk factors of PPCs, including advanced age, higher BMI, lower preoperative serum albumin level, and ICU first day infusion volume.
Advanced age as a predictor of PPCs has been proven by previous studies. Several studies [
18,
19] have found that age > 60 or 65 years is a risk factor for PPCs, which is in agreement with our finding that age was positively correlated with the occurrence of PPCs. In the elderly, elastic fibers around alveoli and capillaries gradually decrease, and the lung tissue elasticity weakens and the retractive ability decreases. Moreover, arteriosclerosis occurs, vascular lumens become thinner, and the number of capillary networks decreases, resulting in reduced pulmonary blood flow, and reduced effective exchange area of the respiratory membrane, all of which greatly weaken lung function in the elderly. However, some studies have also suggested that frailty may be closely related to PPCs in the elderly, and frailty may be a confounding factor between age and PPCs. Future research could further attempt to differentiate frail and non-frail patients in the elderly.
There is mixed evidence about whether higher BMI is a risk of PPCs in the previous literature. Although some studies have shown that obesity conferred a protective factor against PPCs [
20,
21], they stress that BMI does not consider an individual’s body size, and that chest size is more predictive than absolute BMI. In our study, increasing BMI was found to be predictive of PPCs. Obesity causes decreased chest wall compliance, decreased lung volumes, increased oxygen consumption, and increased airway resistance [
22‐
24]. As our study focused on critical patients, the impact of obesity is even more pronounced.
Hypoalbuminemia was also found to be a risk factor for PPCs in our study, which is in agreement with several previous studies [
25‐
27]. The serum albumin level represents the patients’ general nutritional status and their liver reserve function [
28], and albumin is responsible for various biological functions. Decreased albumin concentrations inhibit the activation of macrophages, impair the immune response, and increase susceptibility to infection [
29]. Besides, hypoalbuminemia reduces plasma colloid osmotic pressure, causing pulmonary edema and pleural effusion and making the lungs susceptible to infection [
23,
30].
ICU first day infusion volume in our study means the 24 h infusion volume of the operative day. On one hand, the infusion volume often corelated with the surgery trauma and the hypotension. Larger trauma would cause the severer inflammatory response that would trigger more complications of the patients. And the hypotension is followed by tissue hypoperfusion, which increases postoperative morbidity and mortality [
31,
32]. On the other hand, the infusion volume is associated with the clinician’s cognition and habit of the fluid management and resuscitation. More fluid infusion often associated with the more serious edema that would give rise to the pulmonary edema and the abdominal edema and swelling, eventually, these lead to PPCs. And limited fluid resuscitation has gained more international recognition in recent years [
33]. This study also supports the clinical implementation of restricted fluid resuscitation.
Risk factors are generally divided into two categories: modifiable factors and unmodifiable factors. In our study, the modifiable factors include preoperative serum albumin level and ICU first day infusion while the unmodifiable factors include age and BMI. To reduce the PPCs, a more aggressive human albumin infusion before surgery to elevate the preoperative serum albumin level may be beneficial and the restricted fluid management and resuscitation would help to reduce the PPCs [
34‐
36]. However, all these findings need to be further verified by conducting large-scale prospective randomized controlled trials.
Meanwhile, we create a nomogram to predict the occurrence of PPCs after hepatectomy in critical patients. The nomogram is composed of age, BMI, preoperative serum albumin level and ICU first day infusion volume, and can be fully utilized soon after surgery. When patients with higher probability of PPCs identified by the nomogram, they should be monitored more carefully during the postoperative period. More aggressive lung revascularization, protein supplementation, and fluid restriction may be helpful. To the best of our knowledge, until now there is no study has build a nomogram for PPCs in these patients after hepatectomy especially in the critical patients.
This study has several limitations. First, this is a single-center retrospective study and the lack of reliable information about pulmonary functional tests did not allow precise assessment of the severity of the underlying pulmonary disease. Second, as the sample size of the study was modest, a lack of power is plausible. Third, for our nomogram external validation is required and furthermore, this model would better be developed into both a user-friendly web-based decision aid platform to assist clinicians.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.